The Battle for Open AI’s Soul | WIRED

Overview

Elon Musk’s lawsuit against Sam Altman will head to trial this month in an Oakland, California, federal courtroom, where nine jurors will settle a years-long dispute between the cofounders of Open AI over the group’s founding mission. While spats between Silicon Valley’s most influential billionaires are notable in their own rite, former Open AI employees and nonprofits have taken a special interest in this case because the ruling could influence how the world’s leading AI developer controls and distributes its technology.

The stakes are especially high for Open AI’s corporate future, as a bad outcome in this case could negatively impact its plans to file for an IPO later this year. The Chat GPT-maker is racing against Anthropic and Musk’s Space X (which now owns a rival AI lab, x AI) to go public. Musk’s status as an Open AI competitor—who could benefit significantly if the case goes his way—has raised serious questions about whether he’s the right person to bring it before a jury. A settlement out of court is still possible, though legal experts and people close to the case say it’s unlikely.

Details

Here’s everything you need to know about Musk v. Altman.

Musk’s suit essentially accuses Open AI of straying from its founding nonprofit mission: ensuring AGI, a highly capable AI system that can perform a wide range of jobs, benefits humanity. The defendants in the case are Open AI, Altman, Open AI’s President and cofounder Greg Brockman, and Open AI’s biggest investor, Microsoft.

Despite generating billions of dollars in revenue, Open AI is still overseen by a nonprofit today. Musk was one of the original cofounders of the Open AI nonprofit and donated about $38 million to it during those early days, but he split off in 2018 after getting into disagreements with Altman and Brockman. Now, Musk’s lawsuit has been whittled down to three core claims against Open AI.

The first concerns whether Open AI breached its charitable trust. Musk alleges that in the early days of Open AI, he believed he was investing in a nonprofit with a commitment to open source, or making its AI technology available widely for free download. However, Musk alleges that Altman and Brockman have not used his investment as he intended. Open AI now has a for-profit arm that generates billions of dollars in yearly revenue, and the company is highly secretive about the code for its best AI models. (Open AI alleges that Musk knew back in 2017 that the company would need a for-profit division, and even helped his cofounders set up the corporate structure.)

The second core claim is fraud, and specifically that Altman and Brockman deceived Musk about their intentions to turn Open AI into a for-profit company. The third claim is unjust enrichment, which argues that Altman, Brockman, Microsoft, and other Open AI investors have enriched themselves at the expense of Musk.

The defendants say Musk’s claims are baseless and that he is simply seeking to cripple Open AI as he tries to build up x AI.

Musk is asking the court for a number of different remedies, including removing Altman and Brockman from their roles at Open AI, returning the Chat GPT-maker’s “ill-gotten gains” to the company’s nonprofit, and blocking Open AI from existing as a public benefit corporation, which its for-profit arm is categorized as today.

When reached for comment, an Open AI spokesperson directed WIRED to a section of a company blog that reads, “Motivated by jealousy, regret for walking away from Open AI and a desire to derail a competing AI company, Elon has spent years harassing Open AI through baseless lawsuits and public attacks.” Lawyers for Musk did not respond to multiple requests for comment.

Former Open AI researchers and AI safety nonprofits that have filed amicus briefs in support of Musk in this case say they believe it’s important the Chat GPT maker is held accountable to its founding principles of safety and benefiting humanity, especially as its commercial pressures grow.

Jacob Hilton is part of a group of former Open AI employees that signed one such brief objecting specifically to how Open AI converted into a for-profit entity. “It’s definitely important that Open AI lives up to its mission. I think we’re still seeing a lot of things that Open AI is doing that, in my view, aren’t really consistent with its mission. One recent example people are talking about is backing this Illinois state bill that would shield them from liability,” Hilton says.

Illinois state bill that would shield them from liability

Other groups and experts plan to follow the trial because of its potential to conflict with decisions by the attorneys general of Delaware and California, each of which have regulatory authority over Open AI’s nonprofit. They both already agreed to allow the for-profit conversion to proceed as long as the company follows certain commitments.

The attorneys general “are in a much better position to advocate for the public interest and the charities mission” than Musk, says Nathan Calvin, general counsel for Encode, a nonprofit that has supported several AI safety bills and objected to the Open AI restructuring. He adds that Encode’s focus is on holding Open AI accountable to its AG deals, “particularly given the pressures Open AI may be under to cut corners on those commitments in the context of racing for an IPO.”

Jill Horowitz, law professor with expertise in nonprofits and innovation at Northwestern University, says she is puzzled at why the court is allowing a private actor to challenge a structure that a state authority has just blessed. “It’s not a great precedent for nonprofit law if an aggrieved founder can override the actions of the attorney general,” she says.

Have Any Juicy Details Emerged? Will Any More Come Out?

This lawsuit has already surfaced hundreds of emails between Altman and Open AI’s former chief scientist Ilya Sutskever, entries from Brockman’s diary, and texts between Musk and Mark Zuckerberg—but the trial is likely to reveal a lot more about the core people behind Open AI.

Of course, Musk, Altman, and Brockman will take the stand in this trial. However, several other witnesses from Open AI’s past and present are expected to testify as well. That includes Sutskever, former Open AI CTO Mira Murati, Microsoft CEO Satya Nadella and CTO Kevin Scott, and Open AI Board chairman Bret Taylor. Other central players expected to testify are Shivon Zillis, the former Open AI board member and mother of several of Musk’s children, as well as Jared Birchall, the CEO of Musk’s brain-chip interface company Neuralink and manager of his family office.

It’s possible that former board members involved with Altman’s brief ouster as CEO will testify via video conference, including Helen Toner and Tasha Mc Cauley.

Almost certainly not, but that doesn’t necessarily disqualify it altogether. Luís Calderón Gómez, an associate professor at the Benjamin N. Cardozo School of Law with expertise on tax laws and charities, tells WIRED that Musk may be acting out of self-interest, but he still may be in the right.

“The cynical view is that Musk is trying to hinder Open AI to get [x AI’s AI chatbot] Grok going. But even if that’s right, the right remedy for fraud is to unwind the structure and to have oversight going forward,” said Gómez. “[What Musk is doing:] It’s a good attempt to push back on that general commercialization of charities.”

Hilton, the former Open AI researcher, says he has mixed feelings about the case since Musk “stands to benefit personally.” However, he believes there “are some potentially good outcomes, like the Open AI nonprofit being given more independence, or more light being shed on Open AI’s original mission.”

In your inbox: WIRED's most ambitious, future-defining stories

Big Interview: Chris Hayes makes a living from attention

Livestream replay: Watch our experts discuss Big Tech and the military

Key Takeaways

Elon Musk’s lawsuit against Sam Altman will head to trial this month in an Oakland, California, federal courtroom, where nine jurors will settle a years-long dispute between the cofounders of Open AI over the group’s founding mission
The stakes are especially high for Open AI’s corporate future, as a bad outcome in this case could negatively impact its plans to file for an IPO later this year
Here’s everything you need to know about Musk v
Musk’s suit essentially accuses Open AI of straying from its founding nonprofit mission: ensuring AGI, a highly capable AI system that can perform a wide range of jobs, benefits humanity
Despite generating billions of dollars in revenue, Open AI is still overseen by a nonprofit today